Specialization in a knowledge economy

Key Question Being Explored in This Insight:

Using firm-level data from the US Census Longitudinal Business Database (LBD), Yueyuan Ma discusses novel evidence about a wave of specialization experienced by US firms in the 1980s and 1990s.

View the page of: Yueyuan Ma

Institution: University of California Santa Barbara

Explore more insights in:

Economics Macroeconomics Social Sciences

Media Literacy Toolkit

Faculti's Media Literacy Toolkit helps viewers critically engage with academic insights by analyzing the research context, identifying perspectives, and encouraging thoughtful evaluation.

Critical Questions to Consider ▼

What assumptions does the research make?
Are there alternative perspectives not explored?
What are the limitations of the research method?

Bias and Perspective Awareness ▼

This research comes from University of California Santa Barbara . Reflect on how the institution's academic focus and research partnerships may shape the questions being explored.

Further Reading ▼

Read the Study

Using firm-level data from the US Census Longitudinal Business Database (LBD), Yueyuan Ma discusses novel evidence about a wave of specialization experienced by US firms in the 1980s and 1990s.

Image courtesy of interviewee. August 16, 2024

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month. Login to view 8000+ figures on the platform.

Generating Workspace...

Export as PDF

This video is a detailed presentation of a study on the specialization patterns of U.S. firms during the 1980s and 1990s. The presenter discusses their research using U.S. Census data, highlighting two key trends: firms reduced the number of industries they operated in (shrinkage of production scope), particularly among innovating firms, and a shift towards small firms doing more innovation while large firms focused more on production.

The presenter posits that these changes were driven by improved patent trading markets and stronger patent protections, which were influenced by technological and policy changes. The study introduces a hypothesis suggesting that efficient patent trading reduces mismatches between innovation outputs and a firm’s production capabilities, allowing firms to specialize more effectively and lower management costs.

A structural model was developed to quantify the effects of these changes, concluding that improved patent trading efficiency can explain about 25% of the decline in firms' production scope and about 58% of the shift in innovation activities from large to small firms. This video ultimately contributed to a significant increase in the annual growth rates of the U.S. economy during that period.

The research contributes to discussions on patent trading, patent protection, and the theory of firm boundaries, suggesting that policies should consider the implications of specialization resulting from patent protection. Future research directions include extending the study to a global context, examining how the alignment of patent systems across countries might affect innovation and production choices.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.035 --> 00:01.645
This paper I talk about,

00:01.645 --> 00:03.925
I studied the specialization patterns

00:03.925 --> 00:07.125
of US firms in the 1980s and, and 1990s.

00:07.945 --> 00:10.885
Uh, so in this paper, I used the US census data

00:11.385 --> 00:14.445
and find that US firms experienced a wave

00:14.465 --> 00:17.165
of specialization in the 1980s and 1990s.

00:18.025 --> 00:20.125
And there are two, dimensions.

00:20.575 --> 00:24.245
First of all, I find that US firms decrease the number

00:24.245 --> 00:27.325
of industries in which they produce good and services.

00:28.185 --> 00:30.565
And this is called shrinkage of production scope.

00:31.305 --> 00:34.525
And this shrinkage was more salient for innovating firms.

00:36.105 --> 00:37.405
And second, I find

00:37.405 --> 00:41.645
that in the US innovation production activities became more

00:41.885 --> 00:44.845
separate with small firms doing more innovation

00:45.235 --> 00:47.485
with large firms doing more production.

00:48.425 --> 00:51.645
So this paper asked, what are the driving forces

00:51.905 --> 00:53.845
of these two specialization patterns

00:54.025 --> 00:56.325
and how do they affect economic growth?

00:56.685 --> 00:59.525
I propose a novel hypothesis to explain this,

00:59.525 --> 01:02.605
two specialization patterns that highlight the role

01:03.105 --> 01:06.005
of a more efficient patent trading market

01:06.465 --> 01:08.205
and stronger patent protection.

01:09.085 --> 01:11.145
And those are because of both technology

01:11.145 --> 01:12.545
change and policy changes.

01:13.525 --> 01:15.305
So the hypothesis is the following.

01:16.245 --> 01:18.265
Um, usually there are mismatches

01:18.265 --> 01:21.345
between firm's innovation output and their production.

01:21.815 --> 01:25.305
Like inventors do not usually know whether their innovation

01:25.525 --> 01:29.345
output can be utilized by the firm's production itself.

01:30.285 --> 01:33.945
So when the patent trading market was not like very

01:34.385 --> 01:37.385
efficient, and when the patent was not well protected,

01:38.055 --> 01:41.745
then firm's strategy to reduce this mismatching problem

01:41.775 --> 01:44.985
between innovation production is to try to

01:45.665 --> 01:47.505
increase the production scope such

01:47.505 --> 01:50.945
that their innovation output can be more likely to be

01:51.465 --> 01:53.025
utilized by their own production.

01:54.125 --> 01:57.065
But that will result in a higher management cost.

01:58.595 --> 02:01.455
And then when the patent trading market became more

02:01.895 --> 02:04.555
efficient, then it became easier for firms

02:04.695 --> 02:06.595
to trade patent on the market.

02:06.905 --> 02:08.635
They can buy the patent they want

02:08.695 --> 02:11.275
and also sell the patent that fall outside

02:11.275 --> 02:13.035
of their own production scope.

02:14.595 --> 02:18.775
So then the firms can actually shrink their production scope

02:18.915 --> 02:21.095
and save the management cost.

02:21.515 --> 02:25.655
And that explain why us firms especially innovating once

02:26.295 --> 02:28.895
decrease the number of industries in which they produce good

02:28.915 --> 02:31.535
and services, and that explain the first patent.

02:33.365 --> 02:36.705
And on the other o other hand, small firms,

02:36.705 --> 02:39.105
because they have limited production ability,

02:39.845 --> 02:43.505
so they benefit more from selling patents to the market

02:43.965 --> 02:45.945
and monetize their innovation process.

02:46.725 --> 02:51.065
So small firms have higher incentive to do r and d

02:51.235 --> 02:53.465
after this market became more efficient.

02:54.715 --> 02:58.055
And large firms now can rely more on the market

02:58.195 --> 03:00.125
by buying patents from other firms.

03:01.285 --> 03:04.105
And, they have less incentive to do r and d.

03:04.445 --> 03:08.265
So that's why we observed that large firms decrease their r

03:08.265 --> 03:09.505
and d intensity while small

03:09.505 --> 03:11.265
firms increase their r and d intensity.

03:12.365 --> 03:15.745
So in this paper, I build a structural model to try to

03:16.345 --> 03:19.625
quantify how this more efficient patent treating market

03:19.725 --> 03:22.945
and patent protection affect the specialization patterns

03:22.965 --> 03:24.025
and economic growth.

03:24.155 --> 03:28.345
Build an indulgent growth model with two kind of firms, one

03:28.345 --> 03:30.105
with a higher production ability.

03:30.155 --> 03:31.225
Those are large firms

03:31.605 --> 03:35.065
and one with like, lower production ability.

03:35.065 --> 03:36.265
Those are small firms.

03:37.085 --> 03:40.745
And these two type of firms will indulge choose their

03:40.745 --> 03:44.205
production scope and also their innovation intensity,

03:44.705 --> 03:46.925
taking into account the mismatching

03:46.925 --> 03:49.845
between potential mismatching between their innovation

03:50.425 --> 03:51.885
output and their own production.

03:52.865 --> 03:55.485
And there is a patent trading market,

03:55.825 --> 03:59.045
but the trading efficiency can vary before and

03:59.045 --> 04:00.525
after the policy reform.

04:01.385 --> 04:04.205
And this trading market is over the counter market

04:04.315 --> 04:05.645
with some search frictions.

04:06.105 --> 04:09.245
Before the reform, this search friction was large, and

04:09.255 --> 04:12.205
after the reform, the search friction,

04:12.305 --> 04:14.285
become like, lower.

04:15.025 --> 04:18.045
And then I, I quantitatively analyze

04:18.225 --> 04:23.005
how this decreasing in insert frictions affect firms choice

04:23.145 --> 04:26.245
of production scope and their innovation intensity.

04:26.865 --> 04:30.245
And I find that, this policy

04:30.385 --> 04:33.605
and technology change can explain in the patent trading

04:33.605 --> 04:36.725
market can explain around 25%

04:36.865 --> 04:39.685
of the decrease in firms production scope

04:40.065 --> 04:42.685
and around 58% of the shift

04:42.865 --> 04:45.485
of innovation activities from large to small firms.

04:46.465 --> 04:48.205
And, that result in

04:48.205 --> 04:52.085
around point 64% percentage point increase in the annual

04:52.105 --> 04:53.365
growth rate in the us,

04:53.935 --> 04:56.285
which could be a very large driving force

04:56.345 --> 04:59.925
of economic growth in the US in the 1980s and 1990s.

04:59.955 --> 05:03.645
This paper contribute to, three strand of literature.

05:03.895 --> 05:07.165
First of all, it add to the literature about the impact

05:07.225 --> 05:10.805
of patent trading and patent protection.

05:11.225 --> 05:14.285
So there is a paper by actives

05:14.385 --> 05:16.605
and Greenwood who talk about,

05:16.665 --> 05:20.685
how the patent trade market affect the, similarity

05:20.685 --> 05:23.885
between affirm's innovation and their main line of business.

05:24.625 --> 05:28.045
And this paper move further by considering

05:28.305 --> 05:31.725
how this patent trading market affects firm's boundary

05:31.725 --> 05:34.165
choice, which is a scope choice in this paper.

05:35.025 --> 05:37.205
And also it add to the,

05:37.565 --> 05:39.885
literature about the patent protection

05:40.035 --> 05:43.965
because usually, when people talk about the strengths,

05:44.105 --> 05:47.125
the optimal strengths of patent, patent protection,

05:47.275 --> 05:52.205
usually they talk about the trade off between, the,

05:52.835 --> 05:56.805
like the innovation incentive and, monopoly

05:56.805 --> 05:57.805
Power. Because, you know,

05:57.805 --> 05:59.805
if you have a stronger patent protection,

06:00.295 --> 06:02.365
it'll increase like the incentive

06:02.425 --> 06:04.445
for innovators to do r and d.

06:04.585 --> 06:08.245
But on the other hand, it'll give them, give them loss

06:08.245 --> 06:12.485
of profit if their innovation like, success to succeed.

06:12.585 --> 06:14.965
And that will result in the higher market power.

06:15.705 --> 06:19.925
But this paper consider an often unexplored aspect

06:19.945 --> 06:21.045
of patent protection,

06:21.255 --> 06:23.925
which is the specialization patents driven

06:24.505 --> 06:26.005
by the patent protection.

06:26.305 --> 06:29.605
And that as shown in the paper, actually lead

06:29.605 --> 06:32.205
to a large increase in economic growth.

06:32.205 --> 06:35.325
This paper add to is about the,

06:35.395 --> 06:37.525
like the theory about firm's boundary choice

06:37.835 --> 06:39.325
because there is a, you know,

06:39.325 --> 06:41.045
there is a famous paper by cost.

06:41.545 --> 06:46.365
The cost serum tells us that the internal organization cost

06:46.545 --> 06:49.885
and the external market transaction cost will jointly

06:49.885 --> 06:51.845
determine the boundary of firms.

06:52.545 --> 06:55.795
And here in this paper, I

06:56.435 --> 07:00.195
actually consider the, friction that,

07:00.675 --> 07:02.755
previous paper didn't think, think about,

07:03.045 --> 07:05.075
which is the mismatch problem

07:05.145 --> 07:07.315
between a firm's innovation and production.

07:07.855 --> 07:10.955
And I find that the market that helped

07:11.055 --> 07:12.795
to solve this mismatching problem

07:13.355 --> 07:15.195
actually also shaped firm's boundary.

07:15.335 --> 07:18.195
The main hypothesis, in this model,

07:18.415 --> 07:21.035
or the main assumption of this model is

07:21.035 --> 07:22.835
that there are mismatching problem

07:22.905 --> 07:25.355
between a firm's innovation and production.

07:25.935 --> 07:29.115
So firm's innovation can fall inside

07:29.115 --> 07:30.795
of the firm's own production scope

07:30.975 --> 07:33.515
and then can be utilized by the firm itself.

07:34.145 --> 07:36.195
Well, sometimes it may for outside

07:36.195 --> 07:37.955
of the firm's own production scope,

07:38.575 --> 07:41.115
and then it'll be wasted if you cannot

07:41.505 --> 07:43.075
like, sell it to the market.

07:44.055 --> 07:48.195
And, then I, like in this model, I study

07:48.495 --> 07:50.835
how firms res respond to this mismatch.

07:51.175 --> 07:54.835
And, the patent market will affect, will help

07:54.835 --> 07:56.435
to reduce this mismatching problem.

07:56.655 --> 07:58.555
The patent market efficiency

07:59.055 --> 08:03.235
and the patent, strength of patent protection contribute

08:03.235 --> 08:07.475
to the trading process of, of patents both

08:07.495 --> 08:09.195
for the seller and the buyer.

08:09.735 --> 08:14.195
And that will change firm's choice of their production scope

08:14.375 --> 08:16.395
and their innovation output.

08:16.395 --> 08:19.115
Because there is a more efficient patent trading market,

08:19.565 --> 08:21.755
firms can shrink their production scope

08:21.815 --> 08:23.315
to save the management cost

08:23.315 --> 08:27.035
because they can buy the patent they want from the market

08:27.535 --> 08:30.955
and, sell the patent that fall outside

08:30.955 --> 08:32.355
of their own production scope.

08:33.215 --> 08:35.195
Um, and that

08:35.835 --> 08:39.155
actually explain the first specialization patent

08:39.365 --> 08:42.955
where firms is especially innovating ones decrease the

08:42.955 --> 08:44.595
number of industries in which they

08:44.595 --> 08:45.795
produce good and services.

08:46.775 --> 08:50.955
And in terms of the shift of RD activities between,

08:51.085 --> 08:52.195
large and small firms,

08:52.585 --> 08:53.915
because,

08:53.965 --> 08:56.645
small firms usually have limited production ability,

08:56.995 --> 08:59.045
it's hard for them to produce

08:59.635 --> 09:03.405
mass produce massively produce their own innovation output.

09:03.905 --> 09:06.445
So when the market became more available

09:06.665 --> 09:09.965
or more efficient, their strategy is to innovate

09:10.065 --> 09:13.125
and then sell their innovation output to the market

09:13.465 --> 09:15.645
and monetize their innovation process.

09:16.265 --> 09:19.285
So due to this more efficient market, those small firms

09:19.825 --> 09:22.125
became more incentivized to do r and d

09:22.585 --> 09:24.565
and they increase their r and d intensity.

09:24.745 --> 09:26.245
That's what observe in the data.

09:26.755 --> 09:30.765
Well, on the, on the other side, the big firms now they can,

09:31.065 --> 09:33.285
like rely on the patent market

09:34.025 --> 09:36.725
by buying patents from other firms

09:37.265 --> 09:39.725
so they can focus more on their production process.

09:40.225 --> 09:43.845
That's why in the data I ob I observe that large firms

09:44.805 --> 09:46.845
decrease their r and d intensity will increase

09:46.845 --> 09:47.925
their production intensity.

09:48.435 --> 09:50.605
Usually when people are making these policies,

09:50.755 --> 09:53.885
they talk about the trade off between, the,

09:54.065 --> 09:56.765
the incentives to do r and d and the monopoly power.

09:57.265 --> 09:58.885
But in this paper I find

09:58.885 --> 10:03.085
that the specialization can be a very important result from

10:03.115 --> 10:06.005
this like, like patent protection policies.

10:06.745 --> 10:11.445
Um, and so my, the policy implication of the paper is

10:11.795 --> 10:15.245
that when making the optimal intellectual property rights

10:15.245 --> 10:17.485
protection policies, it'll be helpful

10:17.785 --> 10:21.005
to also take into account of this specialization driven

10:21.105 --> 10:24.565
by the policy and which is often ignored in the like,

10:24.595 --> 10:26.565
current policy making process.

10:28.225 --> 10:32.285
Um, in terms of future research, I think one direction is

10:32.285 --> 10:35.245
to extend the current research to a global setting.

10:35.985 --> 10:39.525
You know, the patent system across different countries has

10:39.545 --> 10:43.365
now become more aligned, which means that the patent market

10:44.285 --> 10:46.085
actually there can be,

10:46.315 --> 10:48.485
like more efficient global patent market

10:49.225 --> 10:52.405
and that will affect the like production scope

10:52.505 --> 10:55.765
and also the innovation intensity choice of different kind

10:55.765 --> 10:57.445
of firms across different countries.

10:58.025 --> 11:01.445
So it'll be interesting to extend this model to a global one

11:01.665 --> 11:04.965
and to talk to, to study how this alignment

11:05.145 --> 11:09.285
of patent system across different countries will affect,

11:09.705 --> 11:11.405
the, the, the innovation

11:11.465 --> 11:13.965
and production choice of like different countries.

Quick Answers

"Exploring Specialization Patterns and Innovation Dynamics of U.S. Firms in the 1980s and 1990s: The Impact of Patent Trading and Policy Reforms"

00:00:00 - This paper I talk about, I studied the specialization patterns of US firms in the 1980s and, and 1990s. so in this paper, I used the US census data and find that US firms experienced a wave of specialization in the 1980s and 1990s. And there are two, dimensions. First of all, I find that US firms decrease the number of industries in which they produce good and services. And this is called shrinkage of production scope. And this shrinkage was more salient for innovating firms. And second, I find that in the US innovation production activities became more
00:00:45 - with large firms doing more production. So this paper asked, what are the driving forces of these two specialization patterns and how do they affect economic growth? I propose a novel hypothesis to explain this, two specialization patterns that highlight the role of a more efficient patent trading market and stronger patent protection. And those are because of both technology change and policy changes. So the hypothesis is the following. usually there are mismatches between firm's innovation output and their production. Like inventors do not usually know whether their innovation
00:01:30 - So when the patent trading market was not like very efficient, and when the patent was not well protected, then firm's strategy to reduce this mismatching problem between innovation production is to try to increase the production scope such that their innovation output can be more likely to be utilized by their own production. But that will result in a higher management cost. And then when the patent trading market became more efficient, then it became easier for firms to trade patent on the market.
00:02:08 - and also sell the patent that fall outside of their own production scope. So then the firms can actually shrink their production scope and save the management cost. And that explain why us firms especially innovating once decrease the number of industries in which they produce good and services, and that explain the first patent. And on the other o other hand, small firms, because they have limited production ability, so they benefit more from selling patents to the market and monetize their innovation process. So small firms have higher incentive to do r and d
00:02:54 - And large firms now can rely more on the market by buying patents from other firms. And, they have less incentive to do r and d. So that's why we observed that large firms decrease their r and d intensity while small firms increase their r and d intensity. So in this paper, I build a structural model to try to quantify how this more efficient patent treating market and patent protection affect the specialization patterns and economic growth. Build an indulgent growth model with two kind of firms, one with a higher production ability. Those are large firms
00:03:35 - Those are small firms. And these two type of firms will indulge choose their production scope and also their innovation intensity, taking into account the mismatching between potential mismatching between their innovation output and their own production. And there is a patent trading market, but the trading efficiency can vary before and after the policy reform. And this trading market is over the counter market with some search frictions. Before the reform, this search friction was large, and after the reform, the search friction,

"Examining the Impact of Patent Market Efficiency on Firm Specialization and Economic Growth in the US (1980s-1990s)"

00:04:15 - And then I, I quantitatively analyze how this decreasing in insert frictions affect firms choice of production scope and their innovation intensity. And I find that, this policy and technology change can explain in the patent trading market can explain around 25% of the decrease in firms production scope and around 58% of the shift of innovation activities from large to small firms. And, that result in around point 64% percentage point increase in the annual growth rate in the us,
00:04:56 - of economic growth in the US in the 1980s and 1990s. This paper contribute to, three strand of literature. First of all, it add to the literature about the impact of patent trading and patent protection. So there is a paper by actives and Greenwood who talk about, how the patent trade market affect the, similarity between affirm's innovation and their main line of business. And this paper move further by considering how this patent trading market affects firm's boundary choice, which is a scope choice in this paper.
00:05:37 - literature about the patent protection because usually, when people talk about the strengths, the optimal strengths of patent, patent protection, usually they talk about the trade off between, the, like the innovation incentive and, monopoly Power. Because, you know, if you have a stronger patent protection, it'll increase like the incentive for innovators to do r and d. But on the other hand, it'll give them, give them loss of profit if their innovation like, success to succeed. And that will result in the higher market power.
00:06:19 - of patent protection, which is the specialization patents driven by the patent protection. And that as shown in the paper, actually lead to a large increase in economic growth. This paper add to is about the, like the theory about firm's boundary choice because there is a, you know, there is a famous paper by cost. The cost serum tells us that the internal organization cost and the external market transaction cost will jointly determine the boundary of firms. And here in this paper, I
00:07:00 - previous paper didn't think, think about, which is the mismatch problem between a firm's innovation and production. And I find that the market that helped to solve this mismatching problem actually also shaped firm's boundary. The main hypothesis, in this model, or the main assumption of this model is that there are mismatching problem between a firm's innovation and production. So firm's innovation can fall inside of the firm's own production scope and then can be utilized by the firm itself. Well, sometimes it may for outside of the firm's own production scope,
00:07:41 - like, sell it to the market. And, then I, like in this model, I study how firms res respond to this mismatch. And, the patent market will affect, will help to reduce this mismatching problem. The patent market efficiency and the patent, strength of patent protection contribute to the trading process of, of patents both for the seller and the buyer. And that will change firm's choice of their production scope and their innovation output. Because there is a more efficient patent trading market,

"Understanding the Impact of Patent Markets on Firm Specialization and R&D Intensity"

00:08:21 - to save the management cost because they can buy the patent they want from the market and, sell the patent that fall outside of their own production scope. and that actually explain the first specialization patent where firms is especially innovating ones decrease the number of industries in which they produce good and services. And in terms of the shift of RD activities between, large and small firms, because, small firms usually have limited production ability, it's hard for them to produce
00:09:03 - So when the market became more available or more efficient, their strategy is to innovate and then sell their innovation output to the market and monetize their innovation process. So due to this more efficient market, those small firms became more incentivized to do r and d and they increase their r and d intensity. That's what observe in the data. Well, on the, on the other side, the big firms now they can, like rely on the patent market by buying patents from other firms so they can focus more on their production process.
00:09:44 - decrease their r and d intensity will increase their production intensity. Usually when people are making these policies, they talk about the trade off between, the, the incentives to do r and d and the monopoly power. But in this paper I find that the specialization can be a very important result from this like, like patent protection policies. and so my, the policy implication of the paper is that when making the optimal intellectual property rights protection policies, it'll be helpful to also take into account of this specialization driven
00:10:24 - current policy making process. in terms of future research, I think one direction is to extend the current research to a global setting. You know, the patent system across different countries has now become more aligned, which means that the patent market actually there can be, like more efficient global patent market and that will affect the like production scope and also the innovation intensity choice of different kind of firms across different countries. So it'll be interesting to extend this model to a global one
00:11:11 - of patent system across different countries will affect, the, the, the innovation and production choice of like different countries.

Faculti AI can make mistakes

Research

Policy

Analysis

Data

Originals

Explainers

Podcasts

New on Faculti

Explainers